errortools is a set of Python tools that facilitate the computation of uncertainty estimates on predictions made by common (machine learning) algorithms.
errortools is a new initiative (start in March 2019) and will be gradually built out. Currently fitting, predicting with, and estimating uncertainties on a logistic regression classifier is being developed, as well as an interface with PyTorch for estimating uncertainties on neural network predictions.
The notebooks folder contains examples on the usage.